The present invention relates to apparatus and methods for converting images recorded on a frame-by-frame basis in cinematic film into field sequential video signals.
Image information recorded on cinematic film typically is converted into video signals by means of the so-called 2-3 pull-down conversion method. With reference to FIG. 1, this technique converts image information from the cinematic film recorded in sequential frames Ka (a=1, 2,...), obtained by "pulling down" the cinematic film at 24 frames per second, to 60 fields of video signals VKn (n=1, 2,...) per second. That is, a pull down mechanism holds each of the sequential frames of the film in a projector gate to alternately permit 2 and 3 field exposures thereof. As illustrated in FIG. 1, from each of the odd numbered frames Ka three identical frames of image data, Km1, Km2 and Km3, where m=1, 3, 5,..., are converted into three respective fields of video signals, VKm1, VKm2 and VKm3. From each of the even numbered frames, two identical frames of image data, Kn1 and Kn2, where n=2, 4, 6,..., are converted into two respective fields of video signals, VKn1 and VKn2.
Accordingly, image information obtained from a first frame K1 of the cinematic film is converted into three fields of video signals VK11, VK12 and VK13, and subsequently image information obtained from a next frame K2 is converted into two fields of video signals VK21 and VK22. Thus, after an image from a first frame of the film is visually displayed for 3/60 of a second, a subsequent image from the second frame is displayed for only 2/60 of a second. Hence, the first image which is displayed for 3/60 of a second provides a stronger visual image than the next image displayed for only 2/60 of a second. As a result, moving objects in the reproduced image produced by such video signals do not appear to move smoothly. Consequently, the video image thus produced suffers from flickering.
In a conventional technique for solving this problem, video signals representing a moving object in the image are extracted, and a displacement of the moving object represented thereby is averaged over time to produce a displacement vector which is fed back to appropriately modify the video signals. It will be appreciated that this technique requires the use of relatively complex, and correspondingly expensive, signal processing equipment for correcting video signals in which a significant number of small objects undergo displacement in differing directions.